Write workspace checkpoints during runs

This commit is contained in:
2026-07-08 02:46:38 +00:00
parent f044c00a7c
commit 1d3a444df8
11 changed files with 1258 additions and 9 deletions

View File

@@ -2,6 +2,8 @@ package cli
import (
"context"
"crypto/sha256"
"encoding/hex"
"encoding/json"
"flag"
"fmt"
@@ -17,6 +19,7 @@ import (
"gitea.maximumdirect.net/eric/notarius/internal/core/config"
"gitea.maximumdirect.net/eric/notarius/internal/core/diagnostics"
"gitea.maximumdirect.net/eric/notarius/internal/core/workspace"
"gitea.maximumdirect.net/eric/notarius/internal/framework/checkpoint"
"gitea.maximumdirect.net/eric/notarius/internal/framework/contracts"
"gitea.maximumdirect.net/eric/notarius/internal/framework/pipeline"
)
@@ -253,6 +256,10 @@ func runPipelineCommand(args []string, stdout, stderr io.Writer, opts Options) i
if err != nil {
return failPipelineCommand(stderr, runDir, cfg.Diagnostics.Retention, fmt.Errorf("create LLM client for profile %q: %w", factoryProfileID, err))
}
checkpointRecorder, err := checkpointRecorderForRun(workspaceSettings, effective.ResolvedPipeline, rawInput, only, llmProfiles, strings.TrimSpace(*llmProfile), strings.TrimSpace(sessionID.value))
if err != nil {
return failPipelineCommand(stderr, runDir, cfg.Diagnostics.Retention, err)
}
output, err := pipeline.New(registries).Run(ctx, pipeline.RunInput{
Pipeline: effective.ResolvedPipeline,
@@ -265,6 +272,7 @@ func runPipelineCommand(args []string, stdout, stderr io.Writer, opts Options) i
LLMProfiles: llmProfiles,
Metadata: runMetadata(*outputDir, *diagnosticsDir),
Warnings: referenceWarnings,
Checkpoints: checkpointRecorder,
})
if err != nil {
if output.Manifest.PipelineID != "" && runDir != nil {
@@ -348,6 +356,68 @@ func writeDiagnostics(runDir *diagnostics.RunDirectory, write func() error) erro
return write()
}
func checkpointRecorderForRun(
settings workspace.Settings,
resolved pipeline.ResolvedPipeline,
rawInput []byte,
only []string,
llmProfiles []artifacts.LLMProfileManifest,
llmProfileOverride string,
sessionID string,
) (pipeline.CheckpointRecorder, error) {
identity, err := workspace.NewCheckpointIdentity(workspace.CheckpointIdentityInput{
Pipeline: resolved,
InputKey: resolved.Input.Module,
RawInputDigest: rawInputDigest(rawInput),
SelectedLanes: only,
RuntimeOverrides: runtimeOverrideFingerprints(llmProfileOverride, sessionID),
References: pipeline.ReferenceProvenance(resolved),
ProvenanceFingerprints: llmProfileFingerprints(llmProfiles),
})
if err != nil {
return nil, fmt.Errorf("create checkpoint identity: %w", err)
}
recorder, err := checkpoint.NewWorkspaceRecorder(settings, identity)
if err != nil {
return nil, fmt.Errorf("create checkpoint recorder: %w", err)
}
return recorder, nil
}
func rawInputDigest(data []byte) string {
sum := sha256.Sum256(data)
return "sha256:" + hex.EncodeToString(sum[:])
}
func runtimeOverrideFingerprints(llmProfileOverride string, sessionID string) []workspace.Fingerprint {
var values []workspace.Fingerprint
if strings.TrimSpace(llmProfileOverride) != "" {
values = append(values, workspace.Fingerprint{Name: "llm_profile_override", Value: strings.TrimSpace(llmProfileOverride)})
}
if strings.TrimSpace(sessionID) != "" {
values = append(values, workspace.Fingerprint{Name: "session_id", Value: strings.TrimSpace(sessionID)})
}
return values
}
func llmProfileFingerprints(profiles []artifacts.LLMProfileManifest) []workspace.Fingerprint {
if len(profiles) == 0 {
return nil
}
values := make([]workspace.Fingerprint, 0, len(profiles))
for _, profile := range profiles {
id := strings.TrimSpace(profile.ID)
if id == "" {
continue
}
values = append(values, workspace.Fingerprint{
Name: "llm_profile:" + id,
Value: strings.TrimSpace(profile.Provider) + ":" + strings.TrimSpace(profile.Model),
})
}
return values
}
func configSource(configPath string) string {
if strings.TrimSpace(configPath) != "" {
return "flag"

View File

@@ -2180,6 +2180,41 @@ func TestRunPipelineWritesWorkspaceDiagnosticsArtifactsOnSuccess(t *testing.T) {
assertPathNotExist(t, filepath.Join(workspaceDir, "debug"))
}
func TestRunPipelineWritesCheckpointsWhenWorkspaceResumeEnabled(t *testing.T) {
workspaceDir := filepath.Join(t.TempDir(), "workspace")
outputDir := t.TempDir()
configPath := writeTestConfig(t, mvpConfigYAMLWithWorkspaceResumeEnabled("dnd-session", workspaceDir, "always"))
inputPath := writeSeriatimInput(t)
var stdout bytes.Buffer
var stderr bytes.Buffer
code := RunWithOptions([]string{"run", "dnd-session", "--config", configPath, "--input", inputPath, "--output-dir", outputDir}, &stdout, &stderr, Options{
LLMClientFactory: fakeLLMFactory(newFakeRunLLMClient(false), nil),
})
if code != 0 {
t.Fatalf("RunWithOptions() code = %d, stderr=%q", code, stderr.String())
}
checkpointDir := onlyCheckpointIdentityDir(t, workspaceDir)
for _, name := range []string{
"source/manifest.json",
"source/source-document.json",
"chunk/manifest.json",
"chunk/chunks.json",
"extract/spells/manifest.json",
"extract/spells/outputs.json",
"merge/spells/manifest.json",
"merge/spells/output.json",
"normalize/spells/manifest.json",
"normalize/spells/output.json",
} {
if _, err := os.Stat(filepath.Join(checkpointDir, name)); err != nil {
t.Fatalf("expected checkpoint artifact %q: %v", name, err)
}
}
assertPathNotExist(t, filepath.Join(workspaceDir, "debug"))
}
func TestRunPipelineSkipsDiagnosticsWhenWorkspaceDiagnosticsDisabled(t *testing.T) {
workspaceDir := filepath.Join(t.TempDir(), "workspace")
outputDir := t.TempDir()
@@ -2335,7 +2370,9 @@ func TestRunPipelineDiagnosticsDirFlagOverridesWorkspaceDiagnosticsOnly(t *testi
t.Fatalf("override diagnostics dir entries = %v, want one run dir", entries)
}
assertPathNotExist(t, filepath.Join(workspaceDir, "diagnostics"))
assertPathNotExist(t, filepath.Join(workspaceDir, "checkpoints"))
if entries := childDirs(t, filepath.Join(workspaceDir, "checkpoints")); len(entries) != 1 {
t.Fatalf("workspace checkpoint pipeline dirs = %v, want one", entries)
}
assertPathNotExist(t, filepath.Join(workspaceDir, "debug"))
}
@@ -2868,6 +2905,24 @@ pipelines:
`
}
func mvpConfigYAMLWithWorkspaceResumeEnabled(pipelineID, workspaceDir, retention string) string {
return `version: 2
workspace:
directory: ` + workspaceDir + `
diagnostics:
enabled: true
retention: ` + retention + `
resume:
enabled: true
pipelines:
` + pipelineID + `:
input: seriatim
artifacts:
spells:
extract: dnd/spells
`
}
func mvpConfigYAMLWithWorkspaceDiagnosticsDisabled(pipelineID, workspaceDir string) string {
return `version: 2
workspace:
@@ -3231,6 +3286,13 @@ func onlyChildDir(t *testing.T, root string) string {
return children[0]
}
func onlyCheckpointIdentityDir(t *testing.T, workspaceDir string) string {
t.Helper()
pipelineDir := onlyChildDir(t, filepath.Join(workspaceDir, "checkpoints"))
inputDir := onlyChildDir(t, pipelineDir)
return onlyChildDir(t, inputDir)
}
func childDirs(t *testing.T, root string) []string {
t.Helper()
entries, err := os.ReadDir(root)